Near Infra Red (NIR) spectroscopy provides a tool to rapidly predict, in a non-destructive way, texture and taste attributes of agro-food products with emphasis on tomatoes and apples. To do so, samples of a tomato or apple variety, varying in texture and taste attributes, are analysed with regard to texture and taste attributes, using both physical, chemical or sensory methods. In addition NIR spectra are also recorded from the intact fruits. Using multivariate statistical techniques, models can be made which relate and predict the defined texture and taste attributes with the NIR spectra. Using these models texture and taste attributes of unknown samples can be predicted. In peaches, the destructive method for mealiness assessment (see result 1) has been used as a reference to check the feasibility of combining NIR and impact techniques in models for non-destructive mealiness assessment.